AI agent that runs in prod
Run serverless AI agents in production: LLM calls, RAG pipelines, and token streaming with AI layers preinstalled. Deploy your first agent in minutes.
LLMRAGStreaming
AI agent backends & toolsFree during beta · No credit card. 10K invocations/mo free.
Serverless backend for AI agents & LLM apps
Deploy Node.js, Python, and Go functions behind one API gateway — long-running, fully traced, with durable retries. No Kubernetes, no servers.
terminallive
$ inquir deploy --template ai-summarizer
✓ Building container 1.2s
✓ Hot runner started 0.3s
✓ Endpoint ready
$ curl -X POST https://app.inquir.org/gw/ws/ai-summarizer \
-d '{"url": "https://techcrunch.com"}'
{
"summary": "AI funding surges as enterprise adoption...",
"wordCount": 824
} 200 OK · 38ms
- 60s
- to first endpoint
- 3
- runtimes (JS, Py, Go)
- Durable
- jobs, retries & DLQ
THE OLD WAY
Shipping one AI feature used to mean five AWS services.
Or — just write the function.
Platform Features
Serverless functions, no infrastructure
One platform for serverless functions, an API gateway, cron schedules, and webhooks. Ship faster than you would wiring up AWS Lambda, run full containers that edge runtimes can't, and never pay the Kubernetes tax.
Fewer cold starts on steady traffic
Hot containers keep your functions warm while traffic keeps flowing, so repeat invocations skip the cold start entirely. The very first request after a deploy may still take the cold path — everything after it stays fast.
Sandboxed by default
Every function runs in its own container with no network access until you grant it. Secrets and environment variables are scoped per function, so one workload can never read another's credentials.
Live HTTPS on every deploy
Every deploy publishes a live HTTPS endpoint automatically — with route groups, path parameters, and response streaming built in. No ingress to configure, no DNS to wait for.
Cron without EventBridge
Schedule cron jobs with a plain cron expression right in the editor. No EventBridge, no external scheduler, no glue code — deploy the function and the schedule ships with it.
Node.js, Python, or Go
Write functions in Node.js 22, Python 3.12, or Go 1.22 and bring any npm or pip package. Full container runtimes mean native modules, headless browsers, and ML libraries just work.
LLM layers preinstalled
OpenAI, Anthropic, and Hugging Face clients come preinstalled as layers, and token streaming to the browser works out of the box — build an LLM feature without packaging a single dependency.
How it works
Deploy in 60 seconds from the browser
Write a serverless function, deploy from the browser, and call a live HTTPS endpoint — no Docker, no config.
Write
Write your handler in the browser editor — Node.js, Python, or Go, with syntax highlighting and AI assistance. No local toolchain to set up.
handler.tsexport default Handler(async (ctx) => { const { url } = await ctx.body; return { summary: await ai.summarize(url) }; });
Deploy
Click Deploy. The platform builds the container, wires the gateway route, and hands you a live HTTPS endpoint — typically in about a second.
terminal$ inquir deploy ✓ Building container 1.2s ✓ Hot runner started 0.3s ✓ Endpoint ready
Call
Call it from anywhere with a standard HTTP request. Any language, plain fetch or curl — no proprietary SDK required.
app.tsxconst res = await fetch(endpoint, { method: "POST", body: JSON.stringify({ url }) });
Watch a live serverless deploy
No login needed. Watch a serverless function build, warm up, and serve a live HTTPS request in real time.
Deploy starts automatically once the code has finished appearing.
More than a function runner
Add gateway routes, scheduled jobs, streaming, and logs without changing your deployment flow.
main.go
package main import ( "context" "encoding/json" "net/http" ) func Handler(ctx context.Context, w http.ResponseWriter, r *http.Request) { var body map[string]any _ = json.NewDecoder(r.Body).Decode(&body) w.Header().Set("Content-Type", "application/json") _ = json.NewEncoder(w).Encode(map[string]any{ "runtime": "go", "echo": body, }) }
Built for LLM
Made for slow, flaky LLM calls
LLM calls are slow, fail sometimes, and are hard to see into. Inquir handles all three — so you ship integrations instead of plumbing.
Long-running
Minute-long model calls and multi-step chains run to completion — no 15-minute edge cutoffs.
Traced end-to-end
Every step's input and output, live run status and logs — no observability to wire up yourself.
Durable & reliable
Background jobs with automatic retries, dead-letter, per-key concurrency and idempotency keys.
Use Cases
Use cases: AI agents, webhooks, cron jobs, REST APIs
From AI agents and webhook processors to background jobs and REST APIs — one platform scales with you, from first deploy to production.
Cron job without EventBridge
Scheduled work without EventBridge or a separate scheduler: built-in cron expressions, automatic retries, and full run history for every tick.
ScheduledSyncETL
Serverless cron jobs guideWebhook handler with retry
Handle Stripe, GitHub, and Slack webhooks reliably: verify signatures, acknowledge fast, and retry idempotently when downstream systems misbehave.
StripeGitHubRetry
Webhook processors & gatewayComparison
Inquir vs AWS Lambda, Cloudflare Workers, Vercel & Modal
APIs, webhooks, cron jobs, and background jobs — without Kubernetes.
Each cell reflects the platform’s core use case. Lambda wins deep inside AWS; Workers wins at the edge; Modal wins for GPU Python. Inquir’s sweet spot: one gateway-first control plane for APIs, webhooks, background jobs, cron, and LLM pipelines.
Feature
Inquir
AWS Lambda
Cloudflare Workers
Vercel
Modal
Container runtime (Node/Py/Go)
(edge isolates)
limited
limited
Built-in cron + schedules
limited
Background jobs + pipelines
limited
limited
Browser IDE + deploy
Pre-built AI layers
limited
Multi-tenant routing built-in
Pricing
Simple serverless pricing
Start free with 10K invocations/mo, then scale with predictable pricing — no compute multipliers.
Free
Get started
$0 during beta- 1 workspace
- 10K invocations / mo
- Hot containers
- API Gateway + cron
Most popular
Starter
For teams
$29 / month- 5 workspaces
- 500K invocations / mo
- Pipelines & webhooks
- Priority support
Pro
No fixed list price
Custom- Unlimited workspaces
- Unlimited invocations
- SLA & dedicated runners
- Custom integrations & SSO
No hidden compute multipliers. Predictable limits. Cancel anytime.
FAQ
Common questions
Yes. The Free tier is free during beta — 10K invocations/mo, hot containers, full API Gateway. No credit card required.
No. Write a function, click Deploy, and get a URL. Inquir builds and runs the containers for you — you never touch a Dockerfile or a cluster.
Inquir runs real containers with full Node.js 22, Python 3.12, or Go 1.22 — any npm or pip package, preinstalled AI layers, and no edge-runtime restrictions on what your code can do.
On Free, extra invocations are rejected gracefully — no surprise bills. On paid plans we notify you before any action.
Each function runs in an isolated container with no network access by default. You configure per-function environment variables in the editor; observability masks common secret-like patterns in logs and traces.
Hi, it's the Inquir Compute team. Ask AI about the product:
Your first live endpoint is 60 seconds away.
Free during beta. No credit card. No Kubernetes.